Keep vanetza-idf in obu-firmware, so a plain clone builds the firmware

obu-firmware builds against the vanetza-idf C-ITS library, which until now
came from the colleague's microbu-esp32c5 tree beside the repository and was
not tracked here, so a clone of this repository could not build the firmware
it ships. The library alone is now part of obu-firmware, as
obu-firmware/external/vanetza-idf: their external/vanetza-idf at commit
cf4b99f, unchanged (9775 files; see its PROVENANCE.md). CMake takes it from
there by default; -DVANETZA_IDF_DIR still points the build elsewhere.

The rest of the colleague's tree (their own VAM firmware, PKI tooling,
station-link Python tools, the V2X2MAP bridge) stays out of this repository
and gitignored; nothing is pushed to their repository. NOTES.md, docs/06,
TODO.md and the pcap verifier's usage line point at the new location.

obu-firmware/external/vanetza-idf/vanetza/common/tests/byte_buffer_source.cpp

@@ -0,0 +1,29 @@
+#include <gtest/gtest.h>
+#include <vanetza/common/byte_buffer_source.hpp>
+#include <boost/iostreams/stream_buffer.hpp>
+#include <istream>
+
+using namespace vanetza;
+
+TEST(ByteBufferSource, read) {
+    const ByteBuffer buf = { 0x01, 0x11, 0x12, 0x22, 0x23, 0x33, 0x34, 0x44 };
+    byte_buffer_source source(buf);
+    boost::iostreams::stream_buffer<byte_buffer_source> stream(source);
+
+    std::istream is(&stream);
+    ASSERT_TRUE(is.good());
+
+    char read = '\0';
+    unsigned read_bytes = 0;
+    for (uint8_t byte : buf) {
+        is >> read;
+        EXPECT_NE(byte, '\0');
+        EXPECT_EQ(byte, read);
+        ++read_bytes;
+    }
+    EXPECT_EQ(read_bytes, 8);
+
+    is >> read;
+    EXPECT_TRUE(is.eof());
+}
+